Introduction

Potato (Solanum tuberosum L.) is the third vegetable food consumed in the world after rice and wheat (Devaux et al. 2014, 2021). Peru is the first potato producer in Latin America and produces 5,331,000 tons of potatoes per year with a productivity of 16.50 t/ha. There are almost 330,000 hectares planted mainly in the highlands with this crop, which improve the resilience of farmers to climate change, contributes to food security, and is the main food for small-scale farmers (MINAGRI 2020). Popular Peruvian potato varieties, known as Canchan, Yungay, and UNICA, occupy more than 50% of the planted area (Pradel et al. 2017). However, they are susceptible to late blight (LB), which is increasing its prevalence due to climate change (Litschmann et al. 2018). The International Potato Center (CIP) has developed potato clones with high levels of resistance to LB, high yields, better agronomic characteristics, and a high potential for release as varieties. They belong to a genetic population known as Population B (Landeo et al. 1995, 1997, 2000; Kerru and Akanda 2005), which in addition to their resistance to LB, have good quality for industrial processing, especially for French fries and chips. These clones can prosper under different agroecological environments (Gonzales et al. 2011; Landeo et al. 2008; Zuñiga et al. 2018; Bajgai et al. 2018; Abebe et al. 2013).

Potato is planted mainly in the Peruvian highlands (90%), where yields are low due to the susceptibility of varieties to LB. Potato production is commercialized, mainly for fresh consumption, and only 4% is processed, especially in French fries. Varieties such as INIA 303-CANCHAN and UNICA (Gastelo et al. 1991; Gutierrez et al. 2007) are currently used for French fries, but they are not stable in their dry matter and reducing sugars content which affects the quality of the frying color and crispness. In 2019, 33,000 tons of frozen pre-fried potatoes were imported from the Netherlands, Germany, and the United States for an approximate value of 30 million dollars (Devaux et al. 2010; MINAGRI 2020; Ordinola and Devaux 2021; Ramos 2022). The potato varieties used in frying industry must have a high percentage of dry matter (> 20%), low content of reducing sugars (< 0.20%), and optimal color of frying (Nawaz et al. 2021; Abong et al. 2009; Naeem and Caliskan 2020; Das et al. 2021; Pumisacho and Sherwood 2000). The dark frying color and unpleasant taste are not accepted by consumers. It occurs when the content of reducing sugars is high, and they are fried at high temperatures, so sugars react with asparagine, an amino acid present in the tubers, producing the "Maillard Reaction” (Wiltshire and Cobb 1996; Pandey et al. 2009) causing the dark fry color.

Dry matter and reducing sugars content depend on the genotype, maturity, temperature, fertilization, and storage of tubers (Kumar et al. 2004). The dry matter content is important because it is correlated with oil absorption. Higher dry matter content is correlated with lower oil absorption (Guido and Mamani 2001). During frying process, amylose and amylopectin present in dry matter promote the formation of a protective layer, preventing the entry of oil at the time of frying (Borruey et al. 2000; Severini et al. 2005).

Reducing sugars content increases with low temperatures to which tubers may be exposed, in the field when they are grown in highlands, where temperatures are low, or when they are stored in a cold chamber until processing (Morales-Fernandez et al. 2015; Hasbún et al. 2009). Reducing sugars have a significant influence on the quality of fried potato since it is highly correlated with dark fry color (Van der Plas 1987; Pritchard and Adam 1994; Moreno 2000). To obtain good frying quality, it is recommended that reducing sugar content should be less than 0.20% of fresh weight (Rodríguez and Wrolstad 1997; Feltran et al. 2004).

These varieties must be stable in yield. AMMI (additive main effects and multiplicative interaction) is a method to determine stability through the main components, the AMMI stability value and yield stability index (Darai et al. 2017; Farshadfar 2008; Farshadfar et al. 2011; Bose et al. 2014; Purchase et al. 2000; Gauch 1992).

The use of these new varieties generated by CIP can be an excellent alternative to face the problem of potato crop in Peru and other countries in Latin America, Africa, and Asia, which have similar problems.

The objective of this study was to obtain potato varieties with high resistance to LB, high yield, and excellent quality for French fries, characteristics that allow the farmers to obtain higher incomes due mainly to the reduction of production costs, increased yield, better price in the market of processing, likewise preserving the environment and human health due to the lower use of pesticides (Islam et al. 2022; Gastelo et al. 2021).

Materials and Methods

Genetic Material Used

Ten potato clones generated by CIP’s genetic improvement program, belonging to genetic population B, subpopulation B3, breeding cycle 1 (4 clones) and breeding cycle 2 (6 clones), were evaluated in the 2019–2020 growing season. Eight clones selected out of ten clones tested initially were evaluated in the 2020–2021 growing season (Table 1). These clones have high levels of resistance to LB, high yield, and high quality for frying in French fries. INIA 303-CANCHAN and UNICA, two popular Peruvian potato varieties known by their susceptibility to LB and their use in processing for French fries, were used as controls.

Table 1 Characteristics of potato clones used in this study
Full size table

Description of Experimental Areas

Thirteen trials were conducted in five regions of Peru (Cajamarca and La Libertad in the north, Huánuco and Junín in the center, and Arequipa in the south): Six (2019–2020 growing season) and seven (2020–2021 growing seasons), all of them conducted at different altitudes, latitudes, temperatures, rainfall, and relative humidity (Table 2).

Table 2 Peruvian localities where the trials were carried out in 2019–2020 and 2020–2021 growing seasons
Full size table

Field Trials and Experimental Design

All experiments were conducted in farmers' fields and implemented with the collaboration of our local partners and collaborators. Experimental plots were arranged in a randomized complete block with three repetitions of 150 plants each, distributed in six rows of 25 plants in each experimental unit. The distance between rows was 1.00 m and between plants 0.30 m. Fertilization was adjusted based on results of soil analysis in each experiment, and a dosage of 200–220–180 kg/hectare of NPK was used with ammonium nitrate (NH4NO3), diammonium phosphate ((NH4)2HPO4) and potassium sulfate (K2SO4) as sources of NPK. Pest control was carried out according to infestation of insects. Only two contact fungicides sprays (Mancozeb) were used in clones, while in control varieties INIA 3030-CANCHAN and UNICA were applied in eight sprays ([Mancozeb + (Propineb + Cymoxanil)]. All trials were planted in the rainy season, except in Arequipa, where sprinkler irrigation was used. The harvest was carried out 120 days after planting. Other agronomic practices were conducted following local customs.

At the harvest, the number of plants harvested and the number and weight of marketable (> 60 g) and non-marketable tubers per experimental unit were registered, manually harvesting the four central rows, leaving one plant unharvested on the edges of the rows to eliminate the border and inter plot effects (27.60 m2), Then marketable tuber yield (MTY) in tons per hectare (t/ha) was calculated. In addition, samples of tubers were used for laboratory tests of dry matter, reducing sugar content, and frying color in French fries.

Determination of Phenotypic Stability for Marketable Tuber Yield (MTY)

Phenotypic stability for marketable tuber yield (MTY) in eight clones and two varieties used as controls were evaluated in two growing seasons using the additive main effect and multiplicative interaction (AMMI) method (Gauch and Zobel 1988; Gauch 1992), which integrates analysis of variance (ANOVA), principal component analysis (PCA), AMMI stability value (ASV) (Purchase et al.2000), and yield stability index (YSI) (Bose et al. 2014); in this study the YSI was identified as MTYSI (marketable tuber yield stability index).

The AMMI model is:

$$Yij=\mu+gi+ej+\Sigma\;\lambda k\;aik\;\gamma jk+\varepsilon ij$$

where:

Yij:

is the yield of the ith genotype in the jth environment.

gi:

is the ith genotype mean deviation.

ej:

is the jth environment mean deviation.

λk:

is the square root of the eigen value of the PC axis k.

αik and  γjk:

are the principal component scores for PC axis k of the ith genotype and the jth environment, respectively

εij:

is the residual

ASV was used as a quantitative stability measure to select genotypes according to their stability for MTY, which was determined using the formula proposed by Purchase et al. (2000). Stable genotypes are those with the lowest ASV value.

$$ASV=\left(\left(Sum\;of\;square\;PC1/Sum\;of\;square\;PC2\right)\left(PC1\;score\right)^2+\left(PC2\;score\right)^2\right)^{1/2}$$

For the selection of stable clones with the highest marketable tuber yield, we used the MTY stability index (MTYSI). Lower values indicate stability and high MTY.

MTYSI was determined using the following formula (Bose et al. 2014):

$$MTYSI= RASV+RMTY$$

where:

RASV:

Rank of the AMMI stability value of clones

RMTY:

Rank of the marketable tuber yield of clones

Validation of Resistance to LB

LB resistance was validated in the 2019–2020 and 2020-2021growing seasons. Trials were installed in endemic places located in Cajamarca and Huanuco, where the temperature ranged between 15 to 20° C, relative humidity with more than 80%, and precipitation with more than 1000 mm per year. LB severity was taken three times between 60 and 80 days after planting, and Area Under Disease Progress Curve (AUDPC) and scale of susceptibility to LB (sAUDPC) were calculated according to Forbes et al. (2014) and Yuen and Forbes (2009). Lower values of AUDPC and sAUDPC indicate more resistance and higher values indicate more susceptibility to LB. The susceptible variety Canchan, with grade 6 on the susceptibility scale, was used as a control.

The AUDPC is a variable that estimates the amount of disease throughout the growing season and is expressed in percentages per day (%/days). The AUDPC is the daily accumulation of infection percent and is directly interpreted without transformation. High values indicate more susceptibility and lower values indicate more resistance (Forbes 2012, Forbes et al. 2014).

The AUDPC per se should not be used to compare potato genotypes between experiments. Furthermore, the AUDPC units as indicators of resistance or susceptibility are not easily interpreted. For example, an AUDPC of 1000 for a genotype planted under severe conditions for the presence of late blight, it can be considered as moderately resistant, while if this same value is obtained under conditions that are not conducive to the development of the disease, it can be considered as susceptible.

To address this problem, Yuen and Forbes (2009) proposed a simple scale (0 high resistant to 9 very susceptible) that is calculated from the AUDPC values; however, to use this scale, it is required to have a susceptible cultivar as a reference in all the experiments that are going to be compared (Forbes et al. 2014; Yuen and Forbes 2009).

Determination of Dry Matter Content

Dry matter content (%) was determined using the hydrometer and oven drying methods (Naeem and Caliskan 2020). The hydrometer method is a non-destructive test, where 3624 g of whole unpeeled potatoes are taken in a basket with the hydrometer, submerged in a container with water, and the specific weight and percentage of dry matter are indicated by the hydrometer. In the oven drying method, 250 g of tubers were cut into small cubes (fresh weight), placed in an oven at 100 °C for 72 h, and then the dry weight was registered. The percentage of dry matter was determined with the following formula:

$$DM\%\;in\;Oven\;drying\;method=\left(Dry\;weight/Fresh\;weight\right)\times100$$

Determination of Reducing Sugars

Reducing sugar content (%) was determined using strips Accu-Chek(R) Active strips. The Accu-Chek Active kit is based on the principle of photometric determination of glucose by staining glucose with oxidoreductases or reaction by glucose dehydrogenase pyrroloquinoline quinone ( PQQ). This kit shows results in quantitative values from 10 mg/dl to 600 mg/dl (Misener et al. 1996; Garcia et al. 2002; Pandey et al. 2009; Brandt 2012). The percentage of reduced sugar was calculated with the following formula:

$$Reducing\;sugar\;percentage=0.000705\;\left(Accu-Chek\;value\right)+0.00453$$

Determination of Frying Color in French Fries

This test was carried out in CIP post-harvest laboratory in La Molina, Lima, Peru (12°4′44''S; 76°55′1''W, 284 m.a.s.l.). The quality of frying color in French fries sticks was measured in strips of 9 mm caliber, and color evaluation of sticks was determined using the USDA French fry color (USDA 1967; Sabbaghi and Ziaiifar 2013). Degrees 1 and 2 indicate an excellent color, degrees 3 and 4 indicate a reasonably good color, and grade 5 is an undesirable dark color (Fig. 1). Three types of frying were carried out, frying at harvest (traditional after the harvest), blanching-frozen and tubers and storage under room conditions for 90 days after the harvest at 14–16 °C (in La Molina, Lima, Peru). These frying methods were compared to see their similarity or difference in the results using Pearson correlation coefficients (α = 0.01) (Wang 2013). The frying process was carried out in two stages: 1) at 160 °C for three minutes and 2) at 180 °C for 2 min using vegetable oil.

Fig. 1
figure 1

USDA French fry color (USDA 1967; Sabbaghi and Ziaiifar 2013)

Full size image

Statistical Analysis

Analysis of variance (ANOVA) for MTY, dry matter and reducing sugars content, and frying color at harvest were performed using the R software version 4.2.1 (R Core Team 2022) and the Agricolae package version 1.3–5 (De Mendiburu F, 2021) and SAS for Windows (Version 9.4 TS Level 1M3). Clone means were discriminated using the least significant difference (LSD, α = 0.05). Combined analysis for these characters was based on responses from eight clones and two control varieties evaluated in two growing seasons.

Pearson correlation coefficients (α = 0.01) (Wang 2013) were used to determine the association between dry matter and reduced sugar content with frying color. Also, it was used to calculate the association between methods used to determine dry matter content: hydrometer and oven drying. Also, there was a calculated association among three types of frying and frying color.

For the selection of clones with potential for varieties were considered the following characters: high resistance to LB, high marketable tuber yield, excellent frying color, high percentage of dry matter, low percentage of reducing sugars.

Data from the experiments were stored and are available in Dataverse (https://doi.org/https://doi.org/10.21223/MXKUIK).

Results

Determination of Phenotypic Stability for Marketable Tuber Yield (MTY)

ANOVA for marketable tuber yield in 13 localities during the 2019–2020 and 2020–2021 growing seasons showed statistically significant differences (P < 0.01) between clones in all localities except in the locality of La Paccha. Clones presented different yields, which means the possibility of selecting superior and stable clones. The coefficients of variation of all experiments ranged between 10.87% to 19.58%.

The marketable yield in the 2019–2020 growing season ranged between 9.63 t/ha with clone CIP391046.14 in Chinchao to 61.58 t/ha with clone CIP396036.201 in Chugay. Based on the overall average of six locations clones, CIP392617.54, CIP392650.12, CIP393077.159, CIP3933371.164, CIP395123.6, CIP396026.101, CIP396034.103, and CIP396036.201 presented higher MTY than control varieties which were also statistically different. Eight clones were selected based on their agronomic performance, resistance to LB, and frying quality, which were tested again in the 2020–2021 growing season. Clone CIP392617.54 was not selected despite having a high yield and resistance to LB due to the presence of tubers with hollow heart (Table 3).

Table 3 Comparison of means for marketable tuber yield within localities (2019–2020 growing season)
Full size table

In the 2020–2021 growing season, the rainfall regime was erratic, with periods of drought and frost, especially in places where the trials depended on rainfall in regions Cajamarca, La Libertad, Huanuco, and Junin. The marketable yield ranged from 10.22 t/ha with clone CIP393077 0.159 to 48.95 t/ha with clone CIP396036.201 in Majes. Considering an overall average of seven localities clones CIP395123.6, CIP396026.101, CIP396034.103, and CIP396036.201 with 32.82, 35.28, 31.24, and 30.58 t/ha, respectively were statistically different from control varieties according to LSD test (P < 0.05 (Table 4). Yields were lower than those obtained in the 2019–2020 growing season due to environmental conditions.

Table 4 Comparison of means for marketable tuber tuber yield within localities (2020–2021 growing season)
Full size table

The AMMI model is appropriate to clarify the effects of genotype x environment interaction, including additive and multiplicative effects (Freeman 1990; Gauch and Zobel 1988; Crossa et al. 1990).

Results of the Combined AMMI ANOVA showed that environments, clones, and clones (genotype) x environments interaction were highly significant (P < 0.01) which are important sources of variation. PC1 and PC2 components explained 36.62% and 23.32% of total Genotype x Environment interaction, which was highly significant (P < 0.01) (Table 5, Fig. 2). This information obtained does not provide a quantitative measure to classify the clones by their phenotypic stability of marketable yield; for this reason, the AMMI stability values (ASV) were estimated according to Purchase et al. (2000). Clones with lower values are considered as most stable.

Table 5 Combined ANOVA AMMI stability for marketable yield for eight potato clones under 13 localities with different environmental conditions (2019–2020 and 2020–2021 growing seasons)
Full size table
Fig. 2
figure 2

Biplot of eight potato clones and thirteen localities with different environmental conditions for marketable tuber yield using PC1 and PC2 scores

Full size image

Clones CIP392650.12, CIP391058.175, CIP395123.6, CIP396026.101, CIP396034.103, and UNICA presented the lowest ASV values for marketable tuber yield through Thirteen contrasting environments in Peru (Table 6). ASV values indicate which clones are stable but not which ones have the highest yield, so we used the MTY stability index (MTYSI) for each clone under study, according to Bose et al. (2014). The lowest MTYSI values indicate stable clones with high tuber yield.

Table 6 Principal components 1 and 2, Marketable tuber yield (MTY), AMMI stability value, and MTY stability index for eight potato clones and two control varieties. Overall average from thirteen localities during the 2019–2020 and 2020–2021 growing seasons
Full size table

Clones CIP395123.6 and CIP396026.101 presented low values of MTYSI, occupying the first places in the MTY ranking in an average of thirteen locations. Similarly, they were in the first places of the ranking in almost all locations; these clones were the most stable and with the highest marketable tuber yield with 38.00 and 36.20 t/ha, respectively, followed by another group formed by clones CIP396034.103, CIP391058.175 and CIP392650.12 with 33.80, 25.00 and 26.40 t/ha respectively. These clones showed that their MTY values were stable across thirteen localities at different altitudes and diverse environmental conditions. Clone CIP396036.201 also showed a high yield with 35.10 t/ha but was not stable due to its high MTYSI value. It showed better adaptation under Majes conditions in the winter-spring season (Tables 6, 7, Fig. 2).

Table 7 Ranking of clones for Marketable tuber yield (MTY) across thirteen localities during the 2019–2020 and 2020–2021 growing seasons
Full size table

Validation of Resistance to LB

ANOVA showed significant statistical differences (P < 0.01) between clones tested in all locations in the 2019–2020 and 2020–2021 growing seasons. This information allowed us to select clones with LB resistance.

All clones except CIP391046.14 showed resistance to LB with AUDPC and sAUDPC values lower than INIA 303-CANCHAN and UNICA, which validated their resistance showed in the previous selection process. Clone CIP395123.6 with average values of AUDPC (62%/days) and sAUDPC (0.29) was the most resistant to LB (Table 8, Fig. 3). The correlation between AUDPC and MTY was negative, with r = -0.68 (Pearson, P < 0.05). A relation between the lower value of AUDPC and higher MTY was obtained.

Table 8 Comparison of means for AUDPC (%-days) and sAUDPC (grades) within localities in 2019–2020 and 2020–2021 growing seasons
Full size table
Fig. 3
figure 3

Overall average resistance to LB in potato clones measured through the AUDPC (2019–2020 and 2020–2021 growing seasons)

Full size image

Determination of Dry Matter Content

Combined ANOVA for dry matter content with hydrometer and oven drying methods showed significant statistical differences (P < 0.01) for environments, clones, and interaction clones x environments (GxE). Results indicated that clones presented different values of dry matter through localities.

In the 2019–2020 growing season, dry matter contents evaluated with the oven drying method ranged from 18.19% with clone CIP393077.159 in La Paccha to 30.38% with clone CIP395123.6 in Quilcas. INIA 303-CANCHAN and UNICA varieties had values ranging from 18.19% to 26.75% (Table 12). In the 2020–2021 growing season, clone CIP393077.159 showed 17.63% dry matter in La Paccha, and it was the only clone with the lowest value. Clone CIP391058.175 obtained the highest value of dry matter in Chugay with 24.16%. Clones CIP391058.175, CIP395123.6, CIP396026.101, and CIP396034.103 were the ones that had dry matter values greater than 20% in all locations in two growing seasons 2019–2020 and 2020–2021 (Table 9 and 10).

Table 9 Comparison of means for dry matter content using oven drying method within locations, 2019–2020 growing season
Full size table
Table 10 Comparison of means for dry matter content using oven drying method within locations (2020-2021growing season)
Full size table

Two methods used to determine dry matter content used in this study presented a statistically significant correlation coefficient (r = 0.79) (Pearson, P < 0.05), which indicated that both methods could be used to estimate dry matter content; however, it is necessary to take care of the hydrometer calibration. On average, the dry matter content obtained with the oven drying method was higher than the results obtained with the hydrometer. This difference is probably due to the hydrometer calibration (Ramos 2011) (Table 11).

Table 11 Overall average dry matter content determined with hydrometer and oven drying methods in all localities tested during the 2019–2020 and 2020–2021 growing seasons
Full size table

Determination of Reducing Sugars Content

Combined ANOVA showed significant statistical differences for sources of variation, environments, clones, and the interaction clones x environments (P < 0.01). These results indicated that environments and clones were different.

The overall average the percentage of reducing sugars obtained in six contrasting locations in the 2019–2020 growing season was less than 0.20%, with statistical differences between clones and control varieties (LSD means comparison test, P < 0.05). Clones CIP391046.14, CIP392617.54, CIP396036.201 and CIP392650.12 obtained more than 0.20% of reducing sugars in localities of La Paccha and Chugay. In contrast, clones CIP393077.159, CIP393371.164, CIP395123.6, CIP396026.101, and CIP396034.103 had less than 0.20% in all locations (Table 12, Fig. 4).

Table 12 Comparison of means for reducing sugars content within localities (2019–2020 growing season)
Full size table
Fig. 4
figure 4

Reducing sugars content (%) in potato clones under thirteen locations during the 2019–2020 and 2020–2021 growing seasons

Full size image

INIA 303-CANCHAN and UNICA varieties obtained reducing sugars contents lower than the maximum permitted for a good frying color quality; however, in localities of Chugay and La Paccha, values obtained by these varieties were higher compared to the other localities (Table 12, Fig. 4).

In the 2020–2021 growing season, clones CIP392650.12, CIP393077.159, CIP 393371.164, and Canchan and UNICA varieties obtained more than 0.20% reducing sugars in Chugay and La Paccha localities. Clone CIP396036.201 presented high percentages of reducing sugars in six of the seven localities (Table 13). In these growing seasons, the average percentage of reduced sugars was higher than in the 2019–2020 growing season due to the presence of periods of drought and frosts that could have influenced reducing sugars content, as reported by Quintana (2018) (Tables  12, 13, Fig. 4).

Table 13 Comparison of percentage of reducing sugars within localities (2020–2021 growing season)
Full size table

Determination of Frying Color in French Fries at Harvest

Combined ANOVA showed statistically significant differences (P < 0.01) for sources of variation: environments, clones, and interaction clones x environments. These results indicated that there were differences between clones and between environments. Response to the color of frying varies from environment to environment due probably to the fact that this characteristic is highly associated with the content of reducing sugars which has high interaction with the environment. In this study, we identified clones with less interaction with the environment or those with grade 1 or 2 on the frying color scale in all locations.

In the 2019–2020 growing season, no differences were found between clones and control varieties in Chinchao, Majes, and Santa Rita. French fries samples of some clones (CIP391046.14, CIP392617.54, CIP392650.12, CIP393077.159, CIP393371.164, and CIP396036.201) clones and INIA 303-CANCHAN and UNICA varieties from Chugay and La Paccha obtained grade 4 of frying color, probably due to the high content of reducing sugars (Table 17). Clones CIP391058.175, CIP395123.6, CIP396026.101, and CIP396034.104 showed degrees 1 or 2 of the frying color scale in samples from all localities (Table 14). Clones CIP391046. 14 and CIP392617.54 were not selected due to their undesirable frying color in Chugay and La Paccha localities and hollow hearts in potato tubers (2020–2021 growing season) (Table 14, Fig. 5).

Table 14 Comparison of frying color in French fries at harvest within localities (2019–2020 growing season)
Full size table
Fig. 5
figure 5

Frying color for French fries in potato clones under thirteen locations during the 2019–2020 and 2020–2021 growing seasons using the USDA scale. USDA 1967; Sabbaghi and Ziaiifar (2013)

Full size image

In the 2020–2021 growing season, statistical differences were found in all locations for frying color between clones and varieties. The highest degrees of fried color were found in clones CIP393077.159, CIP393371.164, CIP396036.201 and INIA 303-CANCHAN and UNICA varieties especially in samples from La Paccha, Chugay, and Yanac due to their high content of reducing sugars obtained in these places. Clones CIP395123.6, CIP396026.101, and CIP396034.103 obtained grades 1 or 2 for frying color in all locations (Table 15).

Table 15 Comparison of means for frying color in French fries at harvest within localities (2020–2021 growing season)
Full size table

To determine the effect of the content of dry matter and reducing sugars on frying color in French fries, a correlation analysis was carried out between these characteristics. A high correlation (P < 0.01) was determined between reducing sugars content and frying color (r = 0.906, R2 = 82%). These results indicate the importance of selecting potato clones with low reducing sugar content in different locations to obtain excellent frying quality (Rodríguez and Wrolstad 1997; Feltran et al. 2004). The correlation between dry matter content and frying color was not significant (r = -0.525, R2 = 28%). In some locations, some clones had more than 20% dry matter but did not show a good frying color due to the high content of reducing sugars, corroborating that this last character is more important than the dry matter content (Fig. 6).

Fig. 6
figure 6

Relation between dry matter, reducing sugar and frying color scale for French fries in potato clones during 2019–2020 and 2020–2021 growing seasons

Full size image

This study used three frying processes: at harvest, with blanching, and in tubers stored under room conditions (up to 90 days after harvest) to determine their shelf life and use in French fries. The correlation between both types of frying process is positive and highly significant (P < 0.01) (Table 16, Fig. 7). Clones CIP395123.6, CIP396026.101, and CIP396034.103 maintained their grade 1 or 2 for frying color in the three types of frying process. It is an excellent alternative for the frying industry, especially when there is no potato harvest. INIA 303-CANCHAN and UNICA also maintained the same frying color with grades 2 and 3 (Table 17).

Table 16 Pearson correlations between frying color at harvest, with blanching, and 90 days after harvest (90 DAH) in potato clones (2019–2020 and 2020–2021 growing seasons)
Full size table
Fig. 7
figure 7

Pearson correlation (P < 0.01) between frying color at harvest with blanching and 90 days after harvest (90 DAH) in potato clones (2019–2020 and 2020–2021 growing seasons)

Full size image
Table 17 Average frying color using USDA scale [USDA (1967); Sabbaghi and Ziaiifar (2013)] over locations for French fries at harvest with blanching and 90 days after harvest in potato clones (2019–2020 and 2020–2021 growing seasons)
Full size table

Discussion

Determination of Phenotypic Stability for Marketable Tuber Yield (MTY)

Potato tuber yield is influenced by genotype x environment interaction; breeding is very important to determine genotypes that can be cultivated as stable varieties in a wide range of contrasting environments and, at the same time, serve as material for breeding programs. In this study, the phenotypic stability of the MTY was determined, using the AMMI method, in thirteen contracting environments located at different altitudes and diverse environmental conditions. The AMMI model is appropriate to clarify the effects of genotype x environment interaction, including additive and multiplicative effects (Freeman 1990; Gauch and Zobel 1988; Crossa et al. 1990).

The principal components information obtained for each clone does not provide a quantitative measure to classify clones by their phenotypic stability. For this reason, the AMMI stability values (ASV) and marketable tuber yield stability index (MTYSI) were estimated. Clones CIP395123.6, CIP396026.101, and CIP396034.103 were the ones that presented the lowest values of ASV and MTYSI, being the most stable and with higher MTY, according to Purchase et al. (2000) and Bose et al. (2014). The control variety, Canchan, was not stable due to its high ASV and MTYSI values, and UNICA, despite having a low ASV value, had a high MTYSI value being stable but with a low MTY. Stable clones can be grown in contrasting environments without affecting their MTY.

Resistance to Late Blight

All clones validated their resistance to LB, presenting lower values of AUDPC and sAUDPC in comparison to Canchan and UNICA, with high values because they are susceptible to this disease. Genetic material with low values of AUDPC and sAUDPC are more resistant, according to Forbes et al. (2014) and Yuen and Forbes (2009). The values of AUDPC and sAUDPC were higher in the 2019–2020 growing season than in the 2020–2021 growing season. In this last season, the rainfall pattern was erratic, probably due to climate change. Likewise, the infection was higher in La Paccha than in Chinchao, probably to different pathogen race compositions in both localities (Gastelo et al. 2022). However, AUDPC values in both seasons were low, which allowed us to corroborate the genetic resistance of the clones estimated before.

Determination of Frying Color and its Components in French Fries

Dry matter and reducing sugar contents (%) are two essential characteristics of having a good frying color (Pandey et al. 2009; Rodríguez and Wrolstad 1997; Feltran et al. 2004). These characters are influenced by genotype x environment interaction. Clones CIP391058.175, CIP395123.6, CIP396026.101 and CIP396034.103 had more than 20% dry matter content in the thirteen locations indicating their high potential to be used in the frying process, according to Pandey et al. (2009). Dry matter contents in clones were higher than in Canchan and UNICA varieties that are currently used for French fries.

To determine the matter content, the hydrometer and oven drying methods were used, which were highly correlated, so both can be used; however, on overall average, dry matter content obtained with the oven drying method was higher than results obtained with the hydrometer. This difference is probably due to the calibration of the hydrometer (Ramos 2011). We suggest that any method can be used, depending on economic resources, bearing in mind that the cost of the oven-drying method is greater than the cost of the hydrometer method and the ease of the process and the quality of the information obtained.

The percentage of reducing sugars is affected by cold temperatures (Hasbún et al. 2009). In places located at altitudes greater than 3000 masl, it is necessary to select clones that are not affected by low temperatures, and they must be stable in their reducing sugars content, according to Morales-Fernandez et al. (2015). In this study, the clones CIP395123.6, CIP396026.101, and CIP396034.103 were identified, which in the thirteen localities maintained their content of reducing sugars below the maximum limit to have a good frying color indicating that these clones were not affected by the environmental conditions in different localities and they have a high potential for use in French fries, taking into account that the percentage of reducing sugars are highly correlated with the frying color, according to Van der Plas (1987), Pritchard and Adam (1994) and Moreno (2000). The clones CIP391046.14, CIP392617.54, CIP396036.201, and CIP392650.12 had more than 0.20% of reducing sugars in localities of La Paccha and Chugay, probably due to the fact that these clones are not stable for this character, which is affected by low temperatures registered in these sites located at higher altitudes according to Rodríguez and Wrolstad (1997) and Feltran et al. (2004), Quintana (2018) and Kizito et al. (2015).

The quality of French fries is determined by the content of reducing sugars, starch, dry matter, and specific weight, blanching, and storage, according to Hedge (2010). Clones CIP395123.6, CIP396026.101, and CIP396034.103 presented grades 1 or 2 for French fries color in all locations in both seasons, probably due to their high dry matter content and low reducing sugar content, according to Khan et al. (2018). When the reducing sugars glucose and fructose are in percentages less than 0.20%, French fries strips are a good color and are acceptable to consumers. The color of the frying is directly related to the content of reducing sugars (Pritchard 1993).

These clones also maintained their good color after being subjected to blanching, a process that reduces reducing sugars to improve frying color (Aguilar et al. 1997), probably because their reduced sugar contents were low, not affecting their flavor and color when fried both traditionally and with blanching, considering the high correlation between reducing sugars and the color of French fries, according to Rodríguez and Wrolstad (1997) and Feltran et al. (2004). Some clones had more than 20% dry matter and did not have a good frying color due to the high content of reducing sugars. Therefore, reducing sugar content is more important than dry matter content, according to what was mentioned by Van der Plas (1987), Pritchard and Adam (1994). and Moreno (2000).

The clones CIP395123.6, CIP396026.101, and CIP396034.103 presented a good frying color with grades 1 and 2 after 90 days of storage under room conditions, probably because their reducing sugar content was not increased since they were not exposed to low temperatures or storage conditions that play a fundamental role in the content of starch and sugars since tuber respiration continues after harvest to maintain some metabolic processes. At high temperatures, the sugar content decreases, and the synthesis of starch increases; the opposite occurs at lower temperatures since the content of reducing sugars increases (Duarte 2019), with a high correlation between the frying color after harvest and 90 days later, being an excellent alternative for the frying industry, especially in times when there is no potato harvest (Valencia 2016).

These results obtained allowed us to select clones CIP395123.6, CIP396026.201, and CIP396034.103 for their excellent quality for French fries, high MTY, high resistance to LB, high content of dry matter and low content of reducing, which are being registered in the corresponding authority in Peru as new varieties, sustainable, healthy and resilient to climate change, which can be incorporated into family farming production systems in Peru and in other countries with similar environmental conditions. These new varieties will allow farmers to give added value to production, obtaining a better price that allows them to improve their living conditions; the lower use of pesticides will allow them to preserve the health of farmers and consumers and conserve the environment; they can also be used as parents in breeding programs.

Conclusions

Marketable tuber yield (MTY) in clones was more significant than in INIA 303-CANCHAN and UNICA varieties. They obtained more than 30 t ha−1 due to their resistance to LB.

Clones CIP395123.6, CIP396026.101 and CIP396034.103 were selected for their stability and high MTY, resistance to LB, dry matter content greater than 20%, low content of reducing sugars less than 0.20%, excellent quality for frying in French fries and long shelf life.

Selected clones have been registered in the official register of cultivars of Peru, as new varieties resilient to climate change, healthy, and sustainable for family farming production systems in Peru and other countries in the world.

Selected clones can be used in breeding programs in developing countries to develop new varieties adapted to their conditions.